Checking the analysis results
Posted: Wed Dec 18, 2024 3:47 am
After entering the above values A and B into Excel, enter your budget in cells G58 and after to calculate the analysis results. The analysis results summarize the expected number of conversions and how much more conversions you can expect by increasing your budget.
Although the slope of the graph becomes slower, the logarithmic trend line is a model that theoretically rises infinitely, so please keep in mind that the accuracy of predictions tends to be low for values in areas where there is no actual data.
(For example, if you have only delivered up to a budget of 2.5 million in the past, but you are simulating a delivery with a budget of 5 million.)
3. Determine your advertising budget based on your phone number lookup saudi arabia company's situation
[Figure 4: Analyzing changes in conversion numbers due to budget fluctuations]
To interpret the analysis results, for example, if the daily budget is 40,000 yen, the number of conversions will be 16.22 and the CPA will be 2,466 yen. If the daily budget is increased to 60,000 yen, the number of conversions will be 18.22 and the CPA will be 3,293 yen. However, since the number of conversions gained by adding 20,000 yen is 2, this is equivalent to using 20,000 yen to acquire 2 conversions. Therefore, the incremental CPA, which takes into account only the net increase in the number of conversions, is evaluated as 10,015 yen.
By increasing or decreasing the daily budget in this way and , you can check how much you can increase the delivery amount and still deliver ads within your company's goals.
Disadvantages of this analysis and cases where it is better to avoid using it because "anyone can easily create it"
Finally, we will introduce cases where the analysis results introduced here are not very useful. As a premise, although the analysis method introduced here is simple, it is highly practical as it is used in actual support situations at our company. However, because it is simple, there are many elements that should be reflected in the analysis results, so we will supplement by explaining in what cases it is better to avoid using this analysis.
Case 1: When the performance data contains many abnormal values or outliers
Since this analysis involves creating a function that can explain the majority of past performance data, it will often not be useful if it contains many values that deviate from the average or outliers.
・ The frequency of sales has increased or decreased compared to previous years, resulting in a large difference in the number of conversions even with the same delivery amount
. ・ The effectiveness of advertising has decreased during periods when inventory has run out.
・ There is a difference in results due to a change in the optimization point of the advertising account.
Case 2: When a major environmental change has occurred between the past and the present
Next, if the fundamental performance of the service or product has changed significantly before and after the analysis, the analysis results will be difficult to use. While the factors explained above are for a specific period, these are factors that have a continuous impact.
・ Consumer behavior has cooled due to COVID-19, causing the CVR of the entire site to drop.
・ Results have changed significantly before and after changing the landing page.
・ TV appearances have increased awareness, making the product/service more likely to be chosen.
Although the slope of the graph becomes slower, the logarithmic trend line is a model that theoretically rises infinitely, so please keep in mind that the accuracy of predictions tends to be low for values in areas where there is no actual data.
(For example, if you have only delivered up to a budget of 2.5 million in the past, but you are simulating a delivery with a budget of 5 million.)
3. Determine your advertising budget based on your phone number lookup saudi arabia company's situation
[Figure 4: Analyzing changes in conversion numbers due to budget fluctuations]
To interpret the analysis results, for example, if the daily budget is 40,000 yen, the number of conversions will be 16.22 and the CPA will be 2,466 yen. If the daily budget is increased to 60,000 yen, the number of conversions will be 18.22 and the CPA will be 3,293 yen. However, since the number of conversions gained by adding 20,000 yen is 2, this is equivalent to using 20,000 yen to acquire 2 conversions. Therefore, the incremental CPA, which takes into account only the net increase in the number of conversions, is evaluated as 10,015 yen.
By increasing or decreasing the daily budget in this way and , you can check how much you can increase the delivery amount and still deliver ads within your company's goals.
Disadvantages of this analysis and cases where it is better to avoid using it because "anyone can easily create it"
Finally, we will introduce cases where the analysis results introduced here are not very useful. As a premise, although the analysis method introduced here is simple, it is highly practical as it is used in actual support situations at our company. However, because it is simple, there are many elements that should be reflected in the analysis results, so we will supplement by explaining in what cases it is better to avoid using this analysis.
Case 1: When the performance data contains many abnormal values or outliers
Since this analysis involves creating a function that can explain the majority of past performance data, it will often not be useful if it contains many values that deviate from the average or outliers.
・ The frequency of sales has increased or decreased compared to previous years, resulting in a large difference in the number of conversions even with the same delivery amount
. ・ The effectiveness of advertising has decreased during periods when inventory has run out.
・ There is a difference in results due to a change in the optimization point of the advertising account.
Case 2: When a major environmental change has occurred between the past and the present
Next, if the fundamental performance of the service or product has changed significantly before and after the analysis, the analysis results will be difficult to use. While the factors explained above are for a specific period, these are factors that have a continuous impact.
・ Consumer behavior has cooled due to COVID-19, causing the CVR of the entire site to drop.
・ Results have changed significantly before and after changing the landing page.
・ TV appearances have increased awareness, making the product/service more likely to be chosen.